1. Field of the Invention
The present invention relates to a processing unit for generating a control signal indicating a quantity of control, a controller with the processing unit for controlling an actuator to give the quantity of control to a controlled object of a vehicle according to the control signal, and a program executed in the processing unit to generate the control signal.
2. Description of Related Art
A controller with a processing unit is disposed on a vehicle to control actuators. Sensors of the vehicle detect physical values as detected parameters, and the processing unit generates control parameters from the detected parameters. Each control parameter indicates a quantity of control for a controlled object. The controller outputs the control parameters to the respective actuators, and each actuator gives a quantity of control to the corresponding controlled object according to the control parameter. To generate the control parameters, the processing unit has recently performed floating-point calculations in addition to fixed-point calculations.
When the processing unit performs both the fixed-point calculation and floating-point calculation, it is required to use fixed-point data and floating-point data in the processing unit. In this case, a volume of calculations is extraordinary increased in the processing unit. To reduce a volume of calculations in the processing unit, inventors of this specification have proposed an electronic control unit having a floating-point calculation function in Published Japanese Patent First Publication No. 2001-195233. This control unit has a floating-point processor in addition to a central processing unit. The floating-point processor performs floating-point calculations to calculate floating-point data from detected parameters and transmits the floating-point data to the control unit. The control unit performs fixed-point calculations to calculate fixed-point data from other detected parameters. Then, the control unit can output the floating-point data to a part of actuators and output the fixed-point data to the other actuators. Therefore, each actuator can give a quantity of control to a controlled object according to the fixed-point data or floating-point data.
However, in this control unit, there is a probability that a failure or fault may occur in the floating-point processor. In this case, even when the control unit is normally operated, a part of actuators receiving the floating-point data cannot correctly give a quantity of control to a controlled object. Therefore, it is required to detect a failure occurring in the floating-point processor.